JP6757303B2 - Breast type identification friend, method and program - Google Patents

Description

The present invention relates to a breast type identification friend, method and program that identifies a breast type based on a breast image obtained by photographing the breast with a mammography imaging apparatus.

In recent years, in order to promote early detection of breast cancer, diagnostic imaging using a radiographic imaging device (called mammography) for photographing the breast has attracted attention. In mammography, the breast is placed on the imaging table and the image is taken while being compressed by the compression plate. The breast is mainly composed of mammary gland tissue and adipose tissue, and it is important for diagnosis to detect lesions such as tumors and calcification hidden in the mammary gland tissue. Therefore, the radiographic image (breast image) of the breast taken by mammography is image-processed by a dedicated operation terminal or the like and then used for diagnosis by a doctor. Doctors check for lesions by displaying a breast image on a display and interpreting it.

On the other hand, the breast is a mixture of mammary gland tissue and adipose tissue, and depending on the ratio of mammary gland tissue to adipose tissue and the distribution of mammary gland tissue and adipose tissue, the breast is classified into high-concentration type, fatty type, scattered mammary gland type and It can be classified into four breast types of non-uniform high concentration type. Of these breast types, breasts classified as high-concentration types have uniform mammary gland tissue, almost no fat mixture, and it is difficult to detect lesions based on breast images. In order to make an accurate diagnosis using the breast image obtained by photographing such a high-density type breast, it is possible to perform image processing to increase the contrast of the high-density type breast breast image. It is desired. For example, in Patent Document 1, image processing such as processing for enhancing contrast with respect to a breast image is performed according to the ratio of mammary gland tissue to adipose tissue in the breast and the distribution state of mammary gland tissue and adipose tissue. A method has been proposed.

Here, regardless of the type of breast, it is conceivable to perform a process of enhancing the contrast on the breast image. However, when the processing for enhancing the contrast is performed on the fatty type breast image, the graininess may be deteriorated or the contrast may be excessively emphasized, resulting in unfavorable image quality for diagnosis. Therefore, in order to perform the process of enhancing the contrast, it is necessary to identify that the breast is a high-density type based on the breast image. However, of the above four types, the high-concentration type and the fatty type all have a single composition of the breast and the images are similar. Therefore, the high-concentration type and the fatty type can be selected by looking at the breast image. It takes skill to identify.

On the other hand, depending on the diagnostic system using the breast image, it may not be possible to acquire the imaging conditions from the mammography imaging apparatus. Therefore, a method for discriminating the type of breast using only breast images has been proposed. For example, in Patent Document 2, a threshold value for the concentration of separating fat and the mammary gland is calculated based on the pixel values of the fat and the breast muscle included in the breast image, and the mammary gland is calculated based on the calculated threshold value. A method of calculating the region and identifying the type of breast based on the ratio of the breast region to the breast region in the breast image has been proposed.

Japanese Unexamined Patent Publication No. 2006-263555 Japanese Unexamined Patent Publication No. 2005-05857

However, the method described in Patent Document 2 requires that the breast be imaged with a positioning that includes the pectoral muscles. Therefore, if the breast image does not include the pectoral muscles, the type of breast cannot be identified.

The present invention has been made in view of the above circumstances, and an object of the present invention is to identify the type of breast by using a breast image containing only the breast.

The breast type identification friend or foe according to the present invention includes a first detection unit that detects a breast region and a skin line from a breast image obtained by photographing the breast with radiation.
A first index value acquisition unit that acquires a first index value indicating the degree of single composition of the breast region,
A second detector that detects the boundary between adipose tissue and mammary gland tissue in a predetermined range from the skin line to the inside of the breast area in the breast image.
A second index value acquisition unit that acquires a second index value indicating the degree of mammary gland clogging with respect to the breast region based on at least one of the strength of the boundary and the distance from the skin line.
It is provided with an identification unit that identifies the type of breast based on the first index value and the second index value.

The "breast region" is a region in which a radiation transmission image of the breast due to radiation transmitted through the breast is represented in the breast image, and the direct radiation region obtained by irradiating the detector directly at the time of imaging is the breast. This is the area excluded from the image.

"Skin line" means the boundary between the skin and the background in the breast image.

The "first index value representing the degree of single composition" is an index value indicating the degree to which the breast region consists of only one composition, and the value becomes larger (or the value becomes larger as the breast region has a single composition. (Becomes smaller).

The "second index value indicating the degree of mammary gland clogging" is an index value indicating how much the mammary gland is contained in the breast region, and the value becomes larger (or smaller) as the breast region contains more mammary glands. It is a thing.

In the breast type identification friend or foe according to the present invention, the first index value acquisition unit may acquire the first index value based on the histogram in the breast region.

Further, in the breast type identification device according to the present invention, the second detection unit generates a line profile at a plurality of positions from the skin line toward the inside of the breast region, and determines the position where the signal value changes with a peak in the line profile. It may be detected as a boundary.

Further, in the breast type identification device according to the present invention, the second index value acquisition unit multiplies the index value representing the strength of the boundary by the weighting coefficient based on the distance from the skin line in each of the plurality of line profiles. Then, the second index value may be obtained by adding the index values representing the strengths of the boundaries multiplied by the weighting coefficients for the plurality of line profiles.

Further, the breast type identification friend or foe according to the present invention may further include a display control unit that displays the breast type on the display unit.

Further, in the breast type identification friend or foe according to the present invention, the display control unit may further display a warning according to the breast type.

The breast type identification method according to the present invention detects a breast region and a skin line from a breast image obtained by photographing the breast with radiation.
Obtain the first index value indicating the degree of single composition of the breast region,
Detects the boundary between adipose tissue and mammary gland tissue in a predetermined range from the skin line to the inside of the breast area in the breast image.
Based on at least one of the strength of the boundary and the distance from the skin line, a second index value indicating the degree of clogging of the mammary gland with respect to the breast region is obtained.
The type of breast is identified based on the first index value and the second index value.

It should be noted that it may be provided as a program for causing a computer to execute the breast type identification method according to the present invention.

The other breast type identification friend or foe according to the present invention has a memory for storing instructions to be executed by a computer, and
The processor comprises a processor configured to execute a stored instruction.
The breast area and skin line are detected from the breast image obtained by photographing the breast with radiation.
Obtain the first index value indicating the degree of single composition of the breast region,
Detects the boundary between adipose tissue and mammary gland tissue in a predetermined range from the skin line to the inside of the breast area in the breast image.
Based on at least one of the strength of the boundary and the distance from the skin line, a second index value indicating the degree of clogging of the mammary gland with respect to the breast region is obtained.
The process of identifying the breast type is executed based on the first index value and the second index value.

According to the present invention, a first index value indicating the degree of single composition of the breast region detected from the breast image is obtained, and at least one of the strength of the boundary between the adipose tissue and the mammary gland tissue and the distance from the skin line. Based on this, a second index value indicating the degree of clogging of the mammary gland with respect to the breast region is obtained, and the type of breast is identified based on the first and second index values. Therefore, the type of breast can be identified even if the breast image does not contain any tissue other than the breast.

Schematic configuration of a radiographic imaging apparatus to which the breast type identification friend or foe according to the first embodiment of the present invention is applied. The view of the radiation imaging apparatus from the direction of arrow A in FIG. The figure which shows the schematic structure of the breast type identification friend or foe according to 1st Embodiment. Diagram to illustrate the detection of breast areas and skin lines Diagram showing histograms according to breast type Diagram to illustrate the approximation of the histogram with a mixed Gaussian distribution Diagram showing the breast sandwiched between the compression plate and the imaging table Diagram for explaining the setting of a predetermined range from the skin line in the breast image Diagram showing multiple lines set in the breast image Diagram showing line profile Diagram showing the boundary between adipose tissue and mammary gland tissue in a breast image The figure which shows the mapping result of the intensity index value and the distance from the boundary skin line The figure which shows the mapping result of the intensity index value and the distance from the boundary skin line Diagram showing a breast image showing the breast type and warning display A flowchart showing the processing performed in the first embodiment The figure which shows the schematic structure of the breast type identification friend or foe by 2nd Embodiment

Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a schematic configuration diagram of a radiation imaging device to which the breast type identification device according to the first embodiment of the present invention is applied, and FIG. 2 is a view of the radiation imaging device viewed from the direction of arrow A in FIG. The radiographic imaging apparatus 1 is a mammography apparatus that photographs a breast M as a subject. As shown in FIG. 1, the radiographic image capturing apparatus 1 includes a photographing unit 10, a computer 2 connected to the photographing unit 10, a display unit 3 connected to the computer 2, and an input unit 4.

The photographing unit 10 includes an arm unit 12. An imaging table 13 is attached to one end of the arm portion 12, and an irradiation unit 14 is attached to the other end so as to face the photographing table 13.

A radiation detector 15 such as a flat panel detector is provided inside the photographing table 13. Further, inside the photographing table 13, a charge amplifier that converts the charge signal read from the radiation detector 15 into a voltage signal, a correlated double sampling circuit that samples the voltage signal output from the charge amplifier, and a voltage signal. A circuit board or the like provided with an AD conversion unit or the like for converting the above into a digital signal is also installed.

The radiation detector 15 can repeatedly record and read a radiation image, and may use a so-called direct type radiation detector that directly receives radiation irradiation to generate a charge, or radiation. You may use a so-called indirect radiation detector that once converts the visible light into visible light and then converts the visible light into a charge signal. Further, as a radiation image signal reading method, a so-called TFT reading method in which a radiation image signal is read by turning on and off a TFT (thin film transistor) switch, or a radiation image signal is generated by irradiating the read light. It is desirable to use a so-called optical reading method that is read, but the present invention is not limited to this, and other things may be used.

An X-ray source 16 which is a radiation source is housed inside the irradiation unit 14. The timing of irradiating X-rays, which is radiation from the X-ray source 16, and the X-ray generation conditions in the X-ray source 16, that is, the imaging conditions such as the tube voltage and the mAs value are controlled by the computer 2.

Further, on the arm portion 12, a compression plate 17 which is arranged above the imaging table 13 and presses and presses the breast M, a support portion 18 which supports the compression plate 17, and a support portion 18 are vertically attached to FIGS. 1 and 2. A moving mechanism 19 for moving in a direction is provided. Information on the height of the compression plate 17, which is the distance between the compression plate 17 and the photographing table 13, is input to the computer 2.

The display unit 3, CRT is (Cathod e Ray Tube) or a display device such as a liquid crystal monitor, other breast image is an X-ray image of the breast M obtained as described below, displays a message or the like necessary for operation To do. The display unit 3 may have a built-in speaker that outputs sound.

The input unit 4 includes a keyboard, a mouse, or a touch panel type input device, and receives an operation of the radiographic image capturing device 1 by an operator. It also accepts instructions for inputting various information such as shooting conditions and correcting the information necessary for shooting. In the present embodiment, each part of the radiation imaging apparatus 1 operates according to the information input from the input unit 4 by the operator.

A breast type identification program is installed on the computer 2. In the present embodiment, the computer may be a workstation or a personal computer directly operated by the operator, or may be a server computer connected to them via a network. The energy subtraction processing program is recorded and distributed on a recording medium such as a DVD (Digital Versatile Disc) or a CD-ROM (Compact Disc Read Only Memory), and is installed in a computer from the recording medium. Alternatively, it is stored in the storage device of the server computer connected to the network or in the network storage in a state where it can be accessed from the outside, and is downloaded and installed in the computer upon request.

FIG. 3 is a diagram showing a schematic configuration of a breast type identification device according to the first embodiment, which is realized by installing a breast type identification program on the computer 2. As shown in FIG. 3, the breast type identification friend or foe according to the first embodiment includes a CPU (Central Processing Unit) 21, a memory 22, and a storage 23 as a standard computer configuration.

The storage 23 is composed of a storage device such as a hard disk or an SSD (Solid State Drive), and stores various information necessary for processing, including a program for driving each part of the radiographic imaging apparatus 1 and a breast type identification program. There is. In addition, the breast image acquired by imaging is also stored.

The memory 22 temporarily stores a breast type identification program or the like stored in the storage 23 in order to cause the CPU 21 to execute various processes. The breast type identification program is a process of causing the CPU 21 to perform X-ray imaging on the radiographic imaging apparatus 1 to acquire a breast image, and a first method of detecting a breast region and a skin line from the breast image. Detection processing, first index value acquisition processing to acquire a first index value indicating the degree of single composition of the breast region, in a predetermined range from the skin line to the inside of the breast region in the breast image, with adipose tissue A second detection process for detecting the boundary with the breast tissue, a second index value for obtaining a second index value indicating the degree of clogging of the breast with respect to the breast region, based on at least one of the strength of the boundary and the distance from the skin line. The index value acquisition process, the identification process for identifying the breast type based on the first index value and the second index value, and the display control process for displaying the breast type identification result on the display unit 3 are defined. There is.

Then, when the CPU 21 executes these processes according to the breast type identification program, the computer 2 has the image acquisition unit 31, the first detection unit 32, the first index value acquisition unit 33, and the second detection unit 34. It functions as a second index value acquisition unit 35, an identification unit 36, and a display control unit 37. The present embodiment is not limited to executing the functions of each part by the software configuration by the breast type identification program, for example, a plurality of ICs (Integrated Circuits), processors, ASICs (Application Specific Integrated Circuits), FPGAs ( The functions of each part may be executed only by the hardware configuration such as Field-Programmable Gate Array), memory, and a combination thereof. Further, the processing of each part may be executed by a combination of the software configuration and the hardware configuration.

The image acquisition unit 31 controls the irradiation unit 14 according to predetermined imaging conditions to acquire the breast image G0. Specifically, the X-ray source 16 is driven to irradiate the breast M with X-rays according to predetermined imaging conditions, and the X-rays transmitted through the breast M are detected by the radiation detector 15 to obtain a breast image. Acquire G0.

The first detection unit 32 detects the breast region and the skin line from the breast image G0. FIG. 4 is a diagram for explaining the detection of the breast region and the skin line. As shown in FIG. 4, the breast image G0 includes the breast region A1 and the direct radiation region A2 corresponding to the region where the radiation detector 15 is directly irradiated with X-rays. Here, the direct radiation region A2 has a higher concentration than the breast region A1. Therefore, the first detection unit 32 detects the breast region A1 from the breast image G0 by performing the threshold value processing using the threshold value for distinguishing the breast region A1 and the direct radiation region A2. Further, the boundary between the detected breast region A1 and the direct radiation region A2 is detected as the skin line SL0.

The first index value acquisition unit 33 acquires the first index value representing the degree of single composition of the breast region A1. The first index value representing the degree of single composition is an index value representing the degree to which the breast region A1 consists of only one composition. For this purpose, the first index value acquisition unit 33 generates a histogram of the breast region A1 detected from the breast image G0. FIG. 5 is a diagram showing a histogram according to the breast type. FIG. 5 shows histograms H1 to H4 for each of the four breast types: high-concentration type, fatty type, interspersed mammary gland type, and heterogeneous high-concentration type.

Here, in the high-concentration type breast, since many mammary glands are contained in the breast region A1, the composition is substantially single. Therefore, the high-concentration type histogram H1 has a single-peak distribution including only one peak. Further, in the fatty type breast, since a large amount of fat is contained in the breast region A1, the composition is substantially single. Therefore, the fatty type histogram H2 also has a monomodal distribution including only one peak.

On the other hand, in the breasts of the mammary gland scattered type and the non-uniform high concentration type, both the mammary gland and the fat are mixed and contained in the breast region A1, and therefore, the breast has a plurality of compositions. Therefore, the histograms H3 and H4 of the mammary gland scattered type and the non-uniform high concentration type have a multi-peaked distribution including a plurality of peaks.

The first index value acquisition unit 33 includes a discriminator for acquiring the degree of single composition of the breast region A1 as the first index value. The discriminator is generated by performing machine learning using a histogram of a large number of single composition breast regions and a histogram of a non-single composition breast region as teacher data. As the machine learning method, any method such as AdaBoost and support vector machine can be used.

In machine learning, the histogram used for learning is approximated by a mixed Gaussian distribution. FIG. 6 is a diagram for explaining an approximation of the histogram with a mixed Gaussian distribution. The histogram H10 having a monomodal distribution shown on the upper side of FIG. 6 is approximated by the Gaussian distribution B11 shown by the broken line, the Gaussian distribution B12 shown by the alternate long and short dash line, and the Gaussian distribution B13 indicated by the alternate long and short dash line. Further, the histogram H20 having a multi-peak distribution shown on the lower side of FIG. 6 is approximated by the Gaussian distribution B21 shown by the broken line, the Gaussian distribution B22 shown by the alternate long and short dash line, and the Gaussian distribution B23 shown by the alternate long and short dash line.

The discriminator takes the parameter of the Gaussian distribution in the histogram, which is the teacher data, as an input, and is trained so that the larger the distribution of the histogram is, the larger the value is output. Therefore, in the present embodiment, the larger the breast region A1 is, the larger the first index value is. Here, as the parameters of the Gaussian distribution, the average value and the variance of the Gaussian distribution are used. When the histogram of the breast region A1 is input to the discriminator trained in this way, the first index value indicating the degree of single composition of the breast region A1 is output. When acquiring the first index value, the input histogram of the breast region A1 is approximated by the mixed Gaussian distribution as in the case of learning, and the Gaussian distribution obtained by approximating the histogram to the discriminator. Parameters are entered.

The second detection unit 34 detects the boundary between the adipose tissue and the mammary gland tissue in a predetermined range from the skin line SL0 toward the inside of the breast region A1 in the breast image G0. Specifically, a line profile is generated at a plurality of positions from the skin line SL0 toward the inside of the breast region A1, and the position where the signal value changes with a peak in the line profile is detected as the boundary between the adipose tissue and the mammary gland tissue.

For this purpose, the second detection unit 34 sets a predetermined range from the skin line SL0 in the breast image G0. FIG. 7 is a diagram showing a breast sandwiched between the compression plate 17 and the imaging table 13, and FIG. 8 is a diagram for explaining the setting of a predetermined range from the skin line SL0 in the breast image G0. At the time of imaging, the breast M is sandwiched between the imaging table 13 and the compression plate 17 as shown in FIG. Therefore, in the breast image G0 acquired by imaging, the region corresponding to the portion of the breast M that is not in contact with the imaging table 13 or the compression plate 17 is formed from the skin line SL0 toward the inside of the breast region A1. Exists.

The second detection unit 34 acquires information on the height of the compression plate 17 from the photographing unit 10. Here, it is assumed that the shape of the breast M is a semicircular cross section in the vicinity of the skin line SL0 of the breast M. In this case, assuming that the height of the compression plate 17 is h, the breast M comes into contact with the compression plate 17 and the imaging table 13 at the point where the distance from the skin line SL0 is h / 2 in the breast region A1 in the breast image G0. Will be. The second detection unit 34 sets a line TL0 composed of a plurality of points whose distance from the skin line SL0 is h / 2. Then, as shown in FIGS. 7 and 8, the region between the skin line SL0 and the line TL0 in the breast region A1 is set to a predetermined range R0 from the skin line SL0 in the breast image G0.

The distance between the skin line SL0 and the line TL0 for setting the predetermined range R0 is statistically determined in advance so that the line TL0 is located in the mammary gland region in the breast region A1. You may use the value.

The second detection unit 34 further sets a plurality of predetermined reference points on the skin line SL0 at equal intervals. The number of reference points may be, for example, 40, but the number is not limited to this. Then, a plurality of lines having each reference point as the start position and the point on the line TL0 closest to each reference point as the end position are set in the breast image G0. FIG. 9 is a diagram showing a line set in the breast image G0. Then, the second detection unit 34 generates a line profile for each line.

The line profile is normalized by the following equation (1) in order to prevent the value from fluctuating due to the thickness of the breast M. Here, Lsi (x, y) is a normalized pixel value on the line Li, and Li (x, y) is a pixel value before normalization on the line Li. The skin line pixel value is the maximum value of the pixel value on the skin line SL0 in the breast image G0. The skin line pixel value is the maximum value that each line Li can take on the line profile.

The mammary gland pixel value is a pixel value of the pixel having the maximum mammary gland content in the breast image G0. Here, for the calculation of the mammary gland content, for example, the method described in JP-A-2010-253245 can be used. In mammography, the method described in JP-A-2010-253245 describes the X-ray dose that directly reaches the radiation detector without penetrating the subject breast, and the X-ray dose that reaches the radiation detector through the breast. This is a method of calculating the mammary gland content based on the relationship between the X-ray attenuation coefficient by fat, the X-ray attenuation coefficient by the mammary gland, and the thickness of the breast. The mammary gland pixel value is the minimum value that each line Li can take on the line profile. FIG. 10 is a diagram showing a line profile.

Lsi (x, y) = (Li (x, y) -mammary gland pixel value) / (skinline pixel value-mammary gland pixel value) (1)

The second detection unit 34 further smoothes each line profile P0. Specifically, the line profile P0 is smoothed by calculating the average value of the pixel values of a plurality of adjacent pixels. The number of pixels for which the average value is calculated can be, for example, 5 pixels, but the number is not limited to this. The smoothed line profile P1 is shown by a alternate long and short dash line in FIG. Then, the second detection unit 34 subtracts the smoothed line profile P1 from the line profile P0 before smoothing to calculate the difference value P2 of the line profile. The difference value P2 of the line profile is shown by a broken line in FIG. As shown in FIG. 10, an upwardly convex peak appears in the difference value P2 of the line profile. The position of this peak corresponds to the position where the signal value changes in the line profile P0. The second detection unit 34 detects the position of the peak of the difference value P2 of the line profile, that is, the position where the signal value of the line profile P0 changes with a peak as the boundary between the adipose tissue and the mammary gland tissue in the breast region A1. To do. By interpolating the boundary detected in each line Li between the lines Li, the boundary line KL0 between the adipose tissue and the mammary gland tissue can be shown in the breast image G0 as shown in FIG.

The second index value acquisition unit 35 acquires a second index value indicating the degree of clogging of the mammary gland with respect to the breast region A1. The second index value is an index value indicating how much the mammary gland is contained, and in the present embodiment, the value becomes larger as the mammary gland is contained more. In order to acquire the second index value, the second index value acquisition unit 35 uses the index value indicating the strength of the boundary between the adipose tissue and the mammary gland tissue detected by the second detection unit 34 and the skin line of the boundary. The distance from SL0 is mapped for each line Li. The index value indicating the strength of the boundary (hereinafter referred to as the strength index value) is the difference value P2 of the line profile calculated when the second detection unit 34 detects the boundary between the adipose tissue and the mammary gland tissue. Use the value of the peak position of. 12 and 13 are diagrams showing the mapping results of the intensity index value and the distance from the boundary skin line SL0. The mapping is represented in a three-dimensional space, but here, for the sake of explanation, the mapping result is shown in a two-dimensional space in which the magnitude of the intensity index value is represented by the size of a circle. Further, in FIGS. 12 and 13, the number of lines is shown up to 20 for the sake of explanation. Here, the larger the intensity index value is, the more clearly the boundary between the adipose tissue and the mammary gland tissue is represented in the breast image G0.

In the mapping shown in FIG. 12, relatively large intensity index values are distributed in a narrow range where the distance from the skin line SL0 is 3 to 5 mm. On the other hand, in the mapping shown in FIG. 13, although a relatively large intensity index value exists, the distance from the skin line SL0 is distributed over a wide range.

The second index value acquisition unit 35 multiplies the strength index value in each line by a weighting coefficient based on the consistency of the distance from the boundary skin line SL0, and adds the strength indexing value multiplied by the weighting coefficient. By doing so, the second index value is acquired. The weighting coefficient based on the consistency of the distance from the skin line is set as follows. That is, the position of the boundary with respect to a certain line Li is compared with the position of the boundary with respect to an adjacent line, for example, the next line Li + 1, and the weighting coefficient is increased so that the closer the position of the boundary with the adjacent line Li + 1 is. To set. For example, if the positions of the boundaries with the adjacent line Li + 1 match, the weighting coefficient is set to 1.0, and the weighting coefficient is reduced as the position of the boundary with the adjacent line Li + 1 is separated.

When the second index value is acquired in this way, the second index value becomes large when the boundary between the adipose tissue and the mammary gland tissue clearly appears in the breast image G0 and is continuous. Here, when the mapping breast image shown in FIG. 12 and the mapping breast image shown in FIG. 13 are compared, the second index value is larger in the mapping breast image shown in FIG.

The identification unit 36 identifies the type of breast based on the first index value and the second index value. Specifically, the first determination is made as to whether or not the first index value is equal to or higher than the predetermined threshold value Th1, and the second index value is equal to or higher than the predetermined threshold value Th2. A second determination is made as to whether or not there is. Here, when the first index value is the threshold value Th1 or more, the degree of single composition is large, so that the breast type can be identified as either a high-concentration type or a fatty type. On the other hand, in the case of the high-concentration type, in the breast region A1 in the breast image G0, the boundary between the fat region and the mammary gland region is clearly and continuously present inside the skin line SL0. Conversely, in the case of the fatty type, there is no clear boundary between the fat region and the mammary gland region. Therefore, when the second index value is compared between the high-concentration type and the fatty type, the high-concentration type has a larger value.

On the other hand, when the first index value is less than the threshold value Th1, the breast type can be identified as either a scattered mammary gland type or a non-uniform high-concentration type. Here, in the case of the non-uniform high-concentration type, in the breast region A1 in the breast image G0, the boundary between the fat region and the mammary gland region exists clearly and continuously inside the skin line SL0. On the contrary, in the case of the scattered mammary gland type, there is no clear boundary between the fat region and the mammary gland region. Therefore, when the second index value is compared between the mammary gland scattered type and the non-uniform high-concentration type, the non-uniform high-concentration type has a larger value.

Therefore, the identification unit 36 identifies the breast type as follows according to the first and second determination results.

First index value ≥ Th1, and second index value ≥ Th2 ... High concentration type

First index value ≥ Th1, and second index value <Th2 ... fatty type

First index value <Th1 and second index value ≧ Th2 ... Non-uniform high concentration type

1st index value <Th1 and 2nd index value <Th2 ... Mammary gland scattered type

The display control unit 37 displays the type of breast identified by the identification unit 36 on the display unit 3. Here, among the four breast types, the high-concentration type tends to overlook the lesion in the image diagnosis, so it is desired to use a method other than the image diagnosis such as ultrasonic examination in combination. Therefore, when the type of breast identified by the identification unit 36 is a high-concentration type, it is preferable to display a warning. As a warning display, for example, it is conceivable to change the color of the characters displayed on the display unit 3, blink the displayed characters, or add a mark only when it is identified as a high density type. Be done. FIG. 14 is a diagram showing the breast type displayed on the display unit 3. As shown in FIG. 14, the identified breast type (high density type in FIG. 14) is displayed in the direct radiation region A2 in the breast image G0. In addition, a warning display 40 with a star is displayed on the left side of the high-concentration type characters.

Next, the processing performed in the present embodiment will be described. FIG. 15 is a flowchart showing the processing performed in the first embodiment. When the input unit 4 receives the instruction to start the process by the operator, the breast M is photographed and the image acquisition unit 31 acquires the breast image G0 (step ST1). Next, the first detection unit 32 detects the breast region A1 and the skin line SL0 from the breast image G0 (step ST2), and the first index value acquisition unit 33 represents the degree of single composition of the breast region A1. The index value of 1 is acquired (step ST3).

Further, the second detection unit 34 detects the boundary between the adipose tissue and the mammary gland tissue in the predetermined range R0 from the skin line SL0 toward the inside of the breast region A1 in the breast image G0 (step ST4). Then, the second index value acquisition unit 35 acquires a second index value indicating the degree of mammary gland clogging with respect to the breast region A1 based on at least one of the strength of the boundary and the distance from the skin line (step ST5). .. Further, the identification unit 36 identifies the breast type based on the first index value and the second index value (step ST6), and the display control unit 37 displays the breast type identification result on the display unit 3. (Step ST7), and the process ends.

As described above, in the present embodiment, the first index value indicating the degree of single composition of the breast region A1 detected from the breast image G0 is obtained, and from the strength of the boundary between the adipose tissue and the mammary gland tissue and the skin line. A second index value indicating the degree of clogging of the mammary gland with respect to the breast region A1 was obtained based on at least one of the distances of the breasts, and the type of breast was identified based on the first and second index values. Therefore, the type of breast can be identified even if the breast image G0 does not include any tissue other than the breast.

Next, a second embodiment of the present invention will be described. FIG. 16 is a diagram showing a schematic configuration of a breast type identification friend or foe according to a second embodiment of the present invention. In FIG. 16, the same reference numbers are assigned to the same configurations as those in FIG. 3, and detailed description thereof will be omitted here. The breast type identification device according to the second embodiment is different from the first embodiment in that it includes an image processing unit 38 that performs image processing on the breast image G0 according to the identified breast type.

Here, in the breast classified into the high concentration type, the mammary gland tissue is uniformly distributed, there is almost no mixture of fat, and it is difficult to detect the lesion based on the breast image G0. Therefore, in the second embodiment, when the breast type is identified as the high density type, the image processing unit 38 performs image processing that emphasizes the contrast with respect to the breast image G0. In this way, when the high density type is identified, the contrast of the breast image G0 can be enhanced by performing image processing for enhancing the contrast of the breast image. By using the breast image G0 with enhanced contrast in this way, it is possible to reduce oversight of lesions.

On the other hand, when the identified breast type is a fatty type, the density may be increased (that is, darkened) in the image processing unit 38. This makes it possible to obtain a breast image G0 having a fat-like image density.

In each of the above embodiments, the formula for calculating the average mammary gland dose, which is the average value of the dose absorbed by the mammary gland for the entire mammary gland, may be changed according to the identified breast type. Here, the average mammary gland dose is calculated by multiplying the irradiation dose by the absorbed dose conversion coefficient. For example, when it is identified as a high concentration type, the average mammary gland dose may be calculated by increasing the value of the absorbed dose conversion coefficient.

Further, in each of the above embodiments, the imaging conditions at the time of the next imaging may be changed according to the identified breast type. Here, at least one of the tube voltage and the mAs value can be used as the imaging condition to be changed. For example, when identified as a type other than the high-concentration type, it is possible to reduce the exposure dose to the breast M while maintaining the image quality of the breast image by reducing at least one of the tube voltage and the mAs value. ..

In each of the above embodiments, the arm portion 12 of the radiation imaging apparatus 1 can be configured so that only the end portion to which the radiation irradiation unit 14 is attached can be rotated. By rotating the arm portion 12 in this way, the radiographic imaging apparatus 1 can perform tomosynthesis imaging and acquire a tomographic image of the breast M. In tomosynthesis imaging, in order to observe the affected area in more detail, the X-ray source is moved to irradiate the subject with radiation from multiple source positions, and the desired tomographic image is taken from the plurality of projected images acquired thereby. This is a method for generating a tomographic image with emphasized surfaces. In tomosynthesis imaging, multiple sources can be moved by moving the X-ray source in parallel with the radiation detector or in a circular or elliptical arc, depending on the characteristics of the imaging device and the required tomographic image. Multiple projected images are acquired by photographing the subject at a position, and these projected images are reconstructed using a reconstruction method such as a shift addition method, a simple back projection method, or a filtered back projection method (FBP method; Filtered Back Projection method). Is reconstructed to generate a tomographic image.

In this way, when the radiographic imaging apparatus 1 is configured to be capable of tomosynthesis imaging, whether or not tomosynthesis imaging is performed following breast M imaging according to the identified breast type in each of the above embodiments. It may be controlled. For example, when the identified breast type is a high concentration type, it is difficult to detect the lesion based on the breast image G0. Therefore, when the identified breast type is a high-density type, the radiographic imaging apparatus 1 may be controlled so that tomosynthesis imaging is performed following imaging of breast M.

Further, in each of the above embodiments, the second index value is acquired based on the strength of the boundary and the distance from the skin line, but based on either the strength of the boundary or the distance from the skin line. , The second index value may be acquired.

Further, in each of the above embodiments, the second detection unit 34 detects the boundary between the adipose tissue and the mammary gland tissue in the breast region A1 based on the difference value of the line profile, but the present invention is limited to this. is not. For example, it is possible to detect the boundary between the adipose tissue and the mammary gland tissue in the breast region A1 by using an arbitrary method such as a method using a differential filter.

1 Radiation imaging device 2 Computer 3 Display unit 4 Input unit 10 Imaging unit 12 Arm unit 13 Imaging table 14 Radiation irradiation unit 15 Radiation detector 16 X-ray source 17 Compression plate 18 Support unit 19 Moving mechanism 21 CPU
22 Memory 23 Storage 31 Image acquisition unit 32 First detection unit 33 First index value acquisition unit 34 Second detection unit 35 Second index value acquisition unit 36 Identification unit 37 Display control unit 38 Image processing unit 40 Warning display A1 Breast area A2 Direct radiation area B11, B12, B13, B21, B22, B23 Gaussian distribution H1-H4, H10, H20 Histogram K0 Boundary line L1, L2, Ln line M Breast G0 Breast image SL0 Skin line TL0 line

Claims (8)

A first detection unit that detects the breast area and skin line from the breast image obtained by photographing the breast with radiation, and
A first index value acquisition unit that acquires a first index value representing the degree of single composition of the breast region, and a first index value acquisition unit.
A second detection unit that detects the boundary between the adipose tissue and the mammary gland tissue in a predetermined range from the skin line to the inside of the breast region in the breast image.
A second index value acquisition unit that acquires a second index value indicating the degree of mammary gland clogging with respect to the breast region based on at least one of the strength of the boundary and the distance from the skin line.
A breast type identification device including an identification unit for identifying the breast type based on the first index value and the second index value. The breast type identification device according to claim 1, wherein the first index value acquisition unit acquires the first index value based on a histogram in the breast region. The second detection unit generates a line profile at a plurality of positions from the skin line toward the inside of the breast region, and detects a position where the signal value changes with a peak in the line profile as the boundary. Or the breast type identification device according to 2. The second index value acquisition unit multiplies the index value representing the strength of the boundary by a weighting coefficient based on the distance from the skin line in each of the plurality of line profiles, and regards the plurality of line profiles. The breast type identification device according to claim 3, wherein an index value representing the strength of the boundary multiplied by the weighting coefficient of the above is added to obtain the second index value. The breast type identification device according to any one of claims 1 to 4, further comprising a display control unit that displays the breast type on the display unit. The breast type identification device according to claim 5, wherein the display control unit further displays a warning according to the breast type. The breast area and skin line are detected from the breast image obtained by photographing the breast with radiation.
Obtain a first index value indicating the degree of single composition of the breast region,
In the breast image, the boundary between the adipose tissue and the mammary gland tissue is detected in a predetermined range from the skin line toward the inside of the breast region.
Based on at least one of the strength of the boundary and the distance from the skin line, a second index value indicating the degree of clogging of the mammary gland with respect to the breast region is obtained.
A breast type identification method for identifying the breast type based on the first index value and the second index value. The procedure for detecting the breast area and skin line from the breast image obtained by photographing the breast with radiation, and
A procedure for obtaining a first index value indicating the degree of single composition of the breast region, and
A procedure for detecting the boundary between adipose tissue and mammary gland tissue in a predetermined range from the skin line to the inside of the breast region in the breast image.
A procedure for obtaining a second index value indicating the degree of mammary gland clogging with respect to the breast region based on at least one of the strength of the boundary and the distance from the skin line.
A breast type identification program that causes a computer to perform a procedure for identifying the breast type based on the first index value and the second index value.